feat: multi-provider AI assistant, repo URL simplification

- lib/ai.ts: unified AI client supporting Anthropic (native format) and
  OpenAI-compatible providers (OpenAI, Perplexity, Ollama, OpenRouter, custom)
- settings: 6-provider chip selector, base URL auto-filled on preset select,
  Anthropic URL hidden (fixed), model/key placeholders per provider
- settings: git repo config replaced by single Repo URL field (.git paste)
  auto-parsed to instance URL + owner + repo at save time
- new.tsx: Suggest button between Symptom and Root Cause, visible when AI
  enabled and title+symptom filled; pre-fills rootCause and fix inline

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

lib/ai.ts

@@ -0,0 +1,128 @@
+import * as SecureStore from "expo-secure-store";
+
+export type AIProvider =
+  | "anthropic"
+  | "openai"
+  | "perplexity"
+  | "ollama"
+  | "openrouter"
+  | "custom";
+
+const KEYS = {
+  AI_PROVIDER: "pref_ai_provider",
+  AI_BASE_URL: "pref_ai_base_url",
+  AI_MODEL: "pref_ai_model",
+  AI_KEY: "pref_ai_key",
+};
+
+export const AI_PROVIDER_BASE_URLS: Record<AIProvider, string> = {
+  anthropic: "https://api.anthropic.com",
+  openai: "https://api.openai.com/v1",
+  perplexity: "https://api.perplexity.ai",
+  ollama: "http://localhost:11434/v1",
+  openrouter: "https://openrouter.ai/api/v1",
+  custom: "",
+};
+
+export const AI_PROVIDER_MODEL_PLACEHOLDERS: Record<AIProvider, string> = {
+  anthropic: "claude-haiku-4-5-20251001",
+  openai: "gpt-4o-mini",
+  perplexity: "llama-3.1-sonar-small-128k-online",
+  ollama: "llama3.2",
+  openrouter: "openai/gpt-4o-mini",
+  custom: "model-name",
+};
+
+const SYSTEM_PROMPT =
+  "You are a systems reliability engineer. Analyze the incident and respond with JSON ONLY " +
+  "(no prose, no markdown fences):\n" +
+  '{"root_cause":"...","fix":"..."}\n' +
+  "Be concise and technical. Max 200 chars per field.";
+
+async function getConfig() {
+  const [provider, baseUrl, model, key] = await Promise.all([
+    SecureStore.getItemAsync(KEYS.AI_PROVIDER),
+    SecureStore.getItemAsync(KEYS.AI_BASE_URL),
+    SecureStore.getItemAsync(KEYS.AI_MODEL),
+    SecureStore.getItemAsync(KEYS.AI_KEY),
+  ]);
+  const p = (provider ?? "openai") as AIProvider;
+  return {
+    provider: p,
+    baseUrl: baseUrl || AI_PROVIDER_BASE_URLS[p],
+    model: model || AI_PROVIDER_MODEL_PLACEHOLDERS[p],
+    key: key ?? "",
+  };
+}
+
+export async function callAI(
+  title: string,
+  symptom: string
+): Promise<{ rootCause: string; fix: string }> {
+  const config = await getConfig();
+  const userMsg = `Title: ${title}\nSymptom: ${symptom}`;
+
+  const raw =
+    config.provider === "anthropic"
+      ? await callAnthropic(config, userMsg)
+      : await callOpenAICompat(config, userMsg);
+
+  try {
+    const parsed = JSON.parse(raw) as { root_cause?: string; fix?: string };
+    return {
+      rootCause: parsed.root_cause ?? "",
+      fix: parsed.fix ?? "",
+    };
+  } catch {
+    throw new Error(`Unexpected AI response: ${raw.slice(0, 120)}`);
+  }
+}
+
+async function callAnthropic(
+  config: { baseUrl: string; model: string; key: string },
+  userMsg: string
+): Promise<string> {
+  const res = await fetch(`${config.baseUrl}/v1/messages`, {
+    method: "POST",
+    headers: {
+      "Content-Type": "application/json",
+      "x-api-key": config.key,
+      "anthropic-version": "2023-06-01",
+    },
+    body: JSON.stringify({
+      model: config.model,
+      max_tokens: 512,
+      system: SYSTEM_PROMPT,
+      messages: [{ role: "user", content: userMsg }],
+    }),
+  });
+  if (!res.ok) throw new Error(`Anthropic ${res.status}: ${await res.text()}`);
+  const data = (await res.json()) as { content: { text: string }[] };
+  return data.content[0].text;
+}
+
+async function callOpenAICompat(
+  config: { baseUrl: string; model: string; key: string },
+  userMsg: string
+): Promise<string> {
+  const headers: Record<string, string> = { "Content-Type": "application/json" };
+  if (config.key) headers["Authorization"] = `Bearer ${config.key}`;
+
+  const res = await fetch(`${config.baseUrl}/chat/completions`, {
+    method: "POST",
+    headers,
+    body: JSON.stringify({
+      model: config.model,
+      max_tokens: 512,
+      messages: [
+        { role: "system", content: SYSTEM_PROMPT },
+        { role: "user", content: userMsg },
+      ],
+    }),
+  });
+  if (!res.ok) throw new Error(`AI ${res.status}: ${await res.text()}`);
+  const data = (await res.json()) as {
+    choices: { message: { content: string } }[];
+  };
+  return data.choices[0].message.content;
+}